[MPS] Fix partitioner parameters recognition

[DenisVieriu97]

Fixes parameters recognition for MPS backend

Testing:

python3 -m unittest backends.apple.mps.test.test_mps -k test_mps_backend_partitioner -v
./cmake-out/examples/apple/mps/mps_executor_runner --model_path mps_backend_partitioner.pte --bundled_program

cc @cccclai, @shoumikhin

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/2777

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 1 Unrelated Failure

As of commit e5fe065 with merge base d4b3e5c ():

NEW FAILURE - The following job has failed:

• Apple / upload-frameworks-ios (gh)
  Credentials could not be loaded, please check your action inputs: Could not load credentials from any providers

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• Apple / test-demo-ios / macos-job (gh)
  RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 1

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[facebook-github-bot]

@cccclai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[cccclai]

Actually, after reading the pr again, why does the following lines mean it's parameter?

[DenisVieriu97]

here by it's parameter we're checking only if it's a constant (lifter buffer)

[cccclai]

it's a constant (lifter buffer)

right, I'm trying to find the logic to determine if it's a constant. This line doesn't look like so.

[DenisVieriu97]

We're checking if it's a constant if is_param or is_buffer are true (or is_get_attr is set to true - which I think it's not the case anymore after moving to lifted graphs)

def is_parameter(exp_prog: torch.export.ExportedProgram, node: torch.fx.Node) -> bool:
    """
    Check if a node is a lifted parameter (static data like weights and bias are
    are supplied as inputs to the graph.

    Args:
        edge_program (torch.export.ExportedProgram): _description_
        node (torch.fx.Node): _description_

    Returns:
        bool: _description_
    """
    return is_get_attr(node) or is_param(exp_prog, node) or is_buffer(exp_prog, node)

[facebook-github-bot]

@cccclai merged this pull request in 6965120.

[cccclai]

@pytorchbot cherry-pick --onto release/0.2 -c critical

[pytorchbot]

Cherry picking #2777

Command git -C /home/runner/work/executorch/executorch cherry-pick -x -X theirs 696512016e01173aadf9b83e2710013bf9de888f returned non-zero exit code 1

The previous cherry-pick is now empty, possibly due to conflict resolution.
If you wish to commit it anyway, use:

    git commit --allow-empty

Otherwise, please use 'git cherry-pick --skip'
On branch cherry-pick-2777-by-pytorch_bot_bot_
You are currently cherry-picking commit 696512016.
  (all conflicts fixed: run "git cherry-pick --continue")
  (use "git cherry-pick --skip" to skip this patch)
  (use "git cherry-pick --abort" to cancel the cherry-pick operation)

nothing to commit, working tree clean

Details for Dev Infra team

Raised by workflow job